Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A light sensor chip, comprising: a light sensor configured to detect light using a first exposure time to output a first image; and a processor electrically connected to the light sensor to receive the first image, and configured to identify an operating mode according to an image parameter associated with the first image, wherein when identifying that the operating mode is a strong light mode, the processor is configured to output the first image, and when identifying that the operating mode is a weak light mode, the processor is configured to convert the first image into a converted image using a pre-stored learning model, and then output the converted image, control the light sensor to detect light using a second exposure time to output a second image, wherein the second exposure time is longer than the first exposure time, and convert, using the pre-stored learning model, the second image into another converted image, and then output the another converted image.
Image sensing technology for improved performance in varying light conditions. This invention relates to a light sensor chip designed to adapt to different lighting environments. The chip includes a light sensor that captures an initial image using a first, shorter exposure time. A processor is connected to the light sensor and analyzes the initial image based on its parameters to determine the current operating mode. If the processor identifies a strong light mode, it simply outputs the initial image. However, if a weak light mode is detected, the processor performs several actions. First, it converts the initial image into a processed image using a pre-stored learning model. Simultaneously, it instructs the light sensor to capture a second image with a second, longer exposure time. This second image is then also processed using the same learning model to generate another processed image. Both the processed initial image and the processed second image are then outputted. This dual-image processing and capture strategy aims to provide better image quality in low-light conditions by leveraging both image enhancement and extended light detection.
2. The light sensor chip as claimed in claim 1 , wherein the learning model is trained by a data network architecture based on a ground truth image.
Light sensor technology for image processing. Addresses the need for improved accuracy and efficiency in light sensor chips by incorporating a learning model trained on ground truth images. The light sensor chip comprises a learning model. This learning model is trained using a data network architecture. The training process utilizes a ground truth image, which serves as a reference for accurate scene representation. This approach enables the sensor chip to interpret captured light information more intelligently, potentially leading to enhanced image quality, reduced noise, or more precise environmental data acquisition. The specific architecture of the data network and the nature of the ground truth image are key to the functionality of this light sensor chip.
3. The light sensor chip as claimed in claim 1 , wherein an image quality, a contrast or a clarity of the converted image is higher than that of the first image, or a blurring of the converted image is lower than that of the first image.
A light sensor chip is designed to capture and process images with improved quality. The chip includes a sensor array that generates a first image from incident light, and a processing circuit that converts the first image into a second image. The processing circuit enhances the second image by improving its image quality, contrast, or clarity compared to the first image, or by reducing blurring. The processing circuit may use techniques such as noise reduction, sharpening, or contrast enhancement to achieve these improvements. The chip may also include additional components like an analog-to-digital converter to digitize the sensor output before processing. The enhanced image output can be used in applications requiring high-quality imaging, such as digital cameras, medical imaging devices, or surveillance systems. The improvements in image quality make the chip suitable for environments with low light or challenging imaging conditions.
4. The light sensor chip as claimed in claim 1 , wherein the image parameter comprises at least one of an image brightness, a gain value, a convergence time of auto exposure and an image quality.
A light sensor chip is designed to enhance image capture performance by dynamically adjusting image parameters based on environmental conditions. The chip includes a light sensor array and a processing unit that analyzes light intensity data from the array to determine optimal settings for image capture. The invention addresses the problem of inconsistent image quality in varying lighting conditions by automatically adjusting parameters such as brightness, gain, auto exposure convergence time, and overall image quality. The processing unit evaluates the light intensity data to detect changes in ambient light and applies corrections to maintain consistent image output. The chip may also include calibration mechanisms to fine-tune sensor sensitivity and reduce noise. By dynamically adjusting these parameters, the light sensor chip ensures high-quality images across different lighting scenarios, improving performance in applications such as digital cameras, smartphones, and surveillance systems. The system may further incorporate machine learning algorithms to predict optimal settings based on historical data, enhancing efficiency and accuracy. The invention provides a robust solution for maintaining image consistency in environments with fluctuating light conditions.
5. The light sensor chip as claimed in claim 1 , wherein the processor does not output the first image in the weak light mode.
A light sensor chip is designed to capture images in varying lighting conditions, including weak light environments. The chip includes a processor that generates image data from sensor signals. In weak light mode, the processor may produce a first image with low signal-to-noise ratio due to insufficient light. To improve image quality, the processor is configured to suppress or omit the output of this first image in weak light conditions. Instead, the processor may apply noise reduction techniques or combine multiple frames to enhance the final output. The chip may also include an image signal processor (ISP) to further refine the image data before output. This approach ensures that only high-quality images are provided to the user, avoiding the display of noisy or unusable images in low-light scenarios. The system dynamically adjusts processing based on ambient light levels, optimizing performance across different environments.
6. The light sensor chip as claimed in claim 1 , wherein when a blurring of the second image is higher than a blur threshold, the processor is further configured to shorten the second exposure time.
A light sensor chip is designed to capture images with improved clarity by dynamically adjusting exposure times based on detected blur. The chip includes an image sensor array for capturing images, a processor for analyzing image blur, and a control unit for adjusting exposure settings. The processor evaluates the blur level of a captured image by comparing it to a predefined blur threshold. If the blur exceeds this threshold, the processor shortens the exposure time for subsequent image captures to reduce motion blur. This adaptive exposure control helps maintain sharpness in images, particularly in scenarios with fast-moving subjects or unstable lighting conditions. The system may also include additional features such as multiple exposure modes, automatic gain control, and noise reduction to enhance image quality further. The dynamic adjustment of exposure time ensures optimal performance across varying environmental conditions, improving the reliability and accuracy of the captured images.
7. An image processing device, comprising: a light sensor chip configured to detect light using a first exposure time to output a first image; and an electronic device coupled to the light sensor chip, and comprising a processor configured to identify an operating mode according to an image parameter associated with the first image, wherein when identifying that the operating mode is a strong light mode, the processor is configured to perform an object identification using the first image, and when identifying that the operating mode is a weak light mode, the processor is configured to convert the first image into a converted image using a pre-stored learning model, and then perform the object identification using the converted image, when an image feature of the first image is lower than a predetermined threshold, the processor is further configured to control the light sensor chip to detect light using a second exposure time to output a second image, wherein the second exposure time is longer than the first exposure time, and convert, using the pre-stored learning model, the second image into another converted image, and then perform the object identification using the another converted image.
This invention relates to an image processing device designed to enhance object identification in varying lighting conditions. The device includes a light sensor chip that captures an initial image using a first exposure time. An electronic device, coupled to the sensor, processes the image based on its quality. If the lighting is strong, the processor directly identifies objects in the initial image. In weak light conditions, the processor applies a pre-stored learning model to enhance the image before performing object identification. If the image quality falls below a predetermined threshold, the processor adjusts the sensor to use a longer second exposure time, capturing a second image. This second image is also processed using the learning model to improve visibility before object identification. The system dynamically adapts exposure times and processing techniques to optimize performance in both bright and low-light environments, ensuring accurate object detection regardless of lighting conditions. The learning model likely employs deep learning or other AI-based techniques to enhance image clarity in weak light scenarios.
8. The image processing device as claimed in claim 7 , wherein the learning model is trained by a data network architecture based on a ground truth image.
This invention relates to image processing devices that use machine learning to enhance image quality. The problem addressed is improving the accuracy and efficiency of image restoration or enhancement tasks, such as denoising, super-resolution, or artifact removal, by leveraging a trained learning model. The device includes a learning model trained using a data network architecture that relies on a ground truth image. The ground truth image serves as a reference for training, allowing the model to learn optimal transformations to improve input images. The learning model is designed to process input images and generate enhanced or restored outputs by applying learned patterns from the training data. The data network architecture may involve convolutional neural networks (CNNs), generative adversarial networks (GANs), or other deep learning frameworks optimized for image processing. The training process adjusts the model's parameters to minimize discrepancies between its predictions and the ground truth image, ensuring high-quality results. This approach improves over traditional image processing methods by automating the enhancement process and adapting to various image degradation scenarios. The trained model can be deployed in real-time applications, such as medical imaging, surveillance, or consumer electronics, where fast and accurate image restoration is critical. The invention focuses on the integration of a robust training framework to ensure the model's effectiveness in practical use.
9. The image processing device as claimed in claim 7 , wherein an image quality, a contrast or a clarity of the converted image is higher than that of the first image, or a blurring of the converted image is lower than that of the first image.
This invention relates to image processing devices designed to enhance image quality. The device processes a first image to generate a converted image with improved visual characteristics. Specifically, the device improves the image quality, contrast, or clarity of the converted image compared to the original first image. Alternatively, the device reduces blurring in the converted image relative to the first image. The processing may involve techniques such as noise reduction, sharpening, or contrast adjustment to achieve these improvements. The device ensures that the converted image is visually superior to the input image, making it suitable for applications requiring high-quality visual output, such as medical imaging, surveillance, or digital photography. The invention addresses the need for automated image enhancement without manual intervention, providing a more efficient and consistent solution for improving image clarity and detail.
10. The image processing device as claimed in claim 7 , wherein the image parameter comprises at least one of an image brightness, a gain value, a convergence time of auto exposure and an image quality.
This invention relates to image processing devices designed to optimize image parameters for improved visual output. The device addresses the challenge of maintaining consistent and high-quality image performance under varying conditions by dynamically adjusting key parameters such as brightness, gain, auto exposure convergence time, and overall image quality. The system includes a processing unit that analyzes input image data and modifies these parameters to enhance the final image. The processing unit may also incorporate feedback mechanisms to refine adjustments based on real-time conditions. Additionally, the device can prioritize certain parameters over others depending on the specific application or user preferences. For example, in low-light scenarios, the system may prioritize brightness and gain adjustments to ensure visibility, while in high-speed imaging, it may focus on reducing auto exposure convergence time to minimize lag. The invention aims to provide a flexible and adaptive solution for optimizing image quality across different environments and use cases.
11. The image processing device as claimed in claim 7 , wherein the processor does not use the first image to perform the object identification in the weak light mode.
The invention relates to an image processing device designed to enhance object identification in weak light conditions. The device includes a processor configured to operate in a weak light mode, where it avoids using a first image for object identification. Instead, the processor relies on alternative techniques to improve detection accuracy in low-light environments. The device may also include an image sensor that captures images under varying lighting conditions and a memory storing reference data for comparison. The processor can analyze image features, apply noise reduction techniques, or utilize additional image frames to compensate for the lack of the first image in weak light mode. This approach helps mitigate the challenges of poor visibility and low signal-to-noise ratios in dim lighting, ensuring more reliable object detection. The device may further include communication interfaces to transmit processed data or receive control signals, enhancing its adaptability in various applications such as surveillance, automotive systems, or industrial automation. The invention aims to improve the robustness of image processing systems in low-light scenarios by dynamically adjusting processing methods based on environmental conditions.
12. The image processing device as claimed in claim 7 , wherein when a blurring of the second image is higher than a blur threshold, the processor is further configured to shorten the second exposure time.
The invention relates to image processing devices designed to improve image quality by dynamically adjusting exposure times based on detected blur levels. The device captures a first image with a first exposure time and a second image with a second exposure time, where the second exposure time is longer than the first. The processor analyzes the second image for blur, comparing it to a predefined blur threshold. If the blur exceeds this threshold, the processor shortens the second exposure time to reduce motion blur in subsequent captures. The device may also apply image processing techniques, such as noise reduction or contrast enhancement, to the first and second images before combining them to produce a final high-quality output. This approach ensures that even in low-light conditions, where longer exposures are typically needed, the system avoids excessive blur by dynamically adjusting exposure parameters. The invention is particularly useful in scenarios where both high sensitivity and sharpness are required, such as in low-light photography or surveillance applications.
13. An operating method of an image processing device, the image processing device comprising a light sensor and a processor coupled to each other, the method comprising: detecting, by the light sensor, light using a first exposure time to output a first image; comparing, by the processor, an image parameter associated with the first image with a parameter threshold; directly using the first image to perform an object identification when the image parameter exceeds the parameter threshold; converting, using a pre-stored learning model, the first image into a converted image and then performing the object identification using the converted image when the image parameter does not exceed the parameter threshold; and when an image feature of the first image is lower than a predetermined threshold, controlling the light sensor to detect light using a second exposure time to output a second image, wherein the second exposure time is longer than the first exposure time, and converting, using the pre-stored learning model, the second image into another converted image, and then performing the object identification using the another converted image.
An image processing device includes a light sensor and a processor that work together to capture and analyze images for object identification. The device addresses challenges in low-light conditions where image quality may be insufficient for accurate object detection. The light sensor captures an initial image using a first exposure time. The processor evaluates an image parameter, such as brightness or contrast, against a predefined threshold. If the parameter exceeds the threshold, the first image is used directly for object identification. If not, the first image is processed through a pre-stored learning model to enhance its quality before object identification. Additionally, if the image features, such as sharpness or detail, fall below a predetermined threshold, the light sensor captures a second image with a longer exposure time to improve clarity. This second image is also processed through the learning model before object identification. The method dynamically adjusts exposure times and applies machine learning-based image conversion to ensure reliable object detection in varying lighting conditions.
14. The operating method as claimed in claim 13 , wherein the learning model is trained by a data network architecture based on a ground truth image.
This invention relates to a method for training a learning model using a data network architecture and a ground truth image. The method addresses the challenge of improving the accuracy and reliability of machine learning models by leveraging high-quality reference data. The learning model is trained using a data network architecture that processes input data and compares it to a ground truth image, which serves as a reference for correct outputs. The ground truth image provides a benchmark for evaluating the model's performance, ensuring that the trained model can accurately replicate or predict the desired outcomes. The data network architecture may include layers such as convolutional, pooling, or fully connected layers, depending on the specific application. The training process involves adjusting the model's parameters to minimize the difference between its predictions and the ground truth image, typically using optimization techniques like gradient descent. This approach enhances the model's ability to generalize from training data to new, unseen inputs, making it suitable for applications in image recognition, medical imaging, autonomous systems, and other fields where precise and reliable predictions are critical. The method ensures that the learning model is trained to produce outputs that closely match the ground truth, improving its overall performance and reliability.
15. The operating method as claimed in claim 13 , wherein an image quality, a contrast or a clarity of the converted image is higher than that of the first image, or a blurring of the converted image is lower than that of the first image.
This invention relates to image processing techniques for enhancing the quality of digital images. The problem addressed is the degradation of image quality in digital images, which can result from factors such as low resolution, poor contrast, or blurring. The invention provides a method for converting a first image into a second, converted image with improved visual characteristics. The method involves processing the first image to generate a converted image that exhibits superior image quality, contrast, or clarity compared to the original. Alternatively, the method reduces blurring in the converted image relative to the first image. The enhancement process may include techniques such as deblurring, contrast adjustment, or resolution improvement. The invention ensures that the converted image retains or improves upon the visual fidelity of the original while mitigating common image defects. This approach is particularly useful in applications requiring high-quality image output, such as medical imaging, surveillance, or digital photography. The method may be implemented in software, hardware, or a combination thereof, and can be applied to various types of digital images, including still images and video frames. The result is a more visually accurate and aesthetically pleasing image suitable for further analysis or display.
16. The operating method as claimed in claim 13 , wherein the image parameter comprises at least one of an image brightness, a gain value, a convergence time of auto exposure and an image quality.
This invention relates to an operating method for optimizing image parameters in a camera system. The method addresses the problem of inconsistent image quality due to varying environmental conditions, such as lighting changes, by dynamically adjusting image parameters to maintain optimal performance. The system captures an image using a camera, processes the image to extract key parameters, and adjusts these parameters in real-time to enhance image quality. The parameters include brightness, gain value, auto exposure convergence time, and overall image quality. The method involves analyzing the captured image to determine the current state of these parameters, then applying adjustments to improve clarity, contrast, and exposure. The adjustments are based on predefined thresholds or algorithms that ensure the image meets desired quality standards. This dynamic adjustment process allows the camera system to adapt to different lighting and environmental conditions without manual intervention, resulting in consistently high-quality images. The invention is particularly useful in applications where real-time image processing is critical, such as surveillance, automotive cameras, and industrial imaging.
17. The operating method as claimed in claim 13 , further comprising: not using the first image to perform the object identification when the image parameter does not exceed the parameter threshold.
This invention relates to image processing systems that perform object identification, particularly in scenarios where image quality or parameters may affect detection accuracy. The problem addressed is the unreliable or inaccurate identification of objects in images when certain image parameters fall below acceptable thresholds, which can lead to false negatives or incorrect classifications. The method involves capturing an image of a scene and analyzing an image parameter, such as resolution, contrast, brightness, or noise level, to determine whether it meets a predefined parameter threshold. If the parameter does not exceed the threshold, the system avoids using the first image for object identification, thereby preventing unreliable results. Instead, the system may wait for a subsequent image with improved parameters or trigger alternative actions, such as adjusting imaging settings or alerting the user. The method ensures that object identification is only performed when image quality is sufficient, improving accuracy and reliability in applications like surveillance, autonomous navigation, or industrial inspection. The system may also include preprocessing steps, such as filtering or enhancement, to improve image quality before parameter evaluation.
Unknown
September 22, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.